The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> SpectreRF vs eldoRF phase noise simulation
https://designers-guide.org/forum/YaBB.pl?num=1088617187

Message started by rfx on Jun 30^th, 2004, 10:39am

Title: SpectreRF vs eldoRF phase noise simulation
Post by rfx on Jun 30^th, 2004, 10:39am

I have simulated phase noise of an LC VCO with spectreRF and eldoRF. The steady state oscillation freuency from both simulators (SST and PSS) are quite close(4346.7MHz and 4346.8MHz). However, phase noise results are quite different;

sim\offset freq 10KHz 100KHz 1MHz
----------------------------------------------------------------
spectreRF -75.8 -99.5 -120.2 (dBc/Hz)
eldoRF -70.4 -96 -117.2 (dBc/Hz)

I have designed VCOs with eldoRF (4 VCOs) before,
and simulated and measured phase noise reults matched within 1-2 dB and also 1/f noise corner were
matching quite close.

Device models are from the same design kit, and I have
checked accuracy of spectreRF phase noise simuation by changing parameters suggested by spectreRF documentation.

Can anybody help me with this discrepency between PN simulation results? Thanks,

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by Paul on Jul 1^st, 2004, 12:42am

Hi,

I'm not an expert in PN simulations, can you say if this is in the 1/f2 or 1/f3 region. Depending on this you will know if it the thermal noise model or the flicker noise model which gives different results.

Paul

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by rfx on Jul 1^st, 2004, 8:44am

Hi Paul,

thanks for reply; there is more discrepence in 1/f3 region
(@10KHz offsett -75.8 versus -70.4 dBc/Hz), and
there is less discerpency in 1/f2 region
(@1MHz offsett -120.2 versus -177.2 dBc/Hz).

So there are problems in both 1/f3 and 1/f3 regions.

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by Andrew Beckett on Jul 1^st, 2004, 9:45pm

I assume you have set maxsidebands to something reasonable (not 0)?

There was a mistake in some earlier (up until IC446, I think) documentation which suggested setting maxsidebands to 0 for oscillators.

Regards,

Andrew.

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by rfx on Jul 2^nd, 2004, 7:59am

Hi Andrew,

Thanks for kind reply. I have set maxsidebands to 7 initially, and then also set to 15, 30 to see any difference.

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by Paul on Jul 11^th, 2004, 12:11pm

Hi,

although the models are from the same design kit, there may be some default settings which differ between SpectreRF and EldoRF. It seems the thermal noise model is not computed the same. You should go through both simulator documentations and find out the equations for thermal noise calculation.

I noticed your VCO is running at 4.3GHz. I expect you are using minimum length devices. Check your noise models if there is anything about thermal noise increase due to short channel effects. This may be modeled differently in both cases.

Paul

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by rfx on Jul 13^th, 2004, 7:43am

Hi Paul,

Thanks for comments and suggestions.

I have investigated little bit more since my last posting.
I totally agree with your findings. There are two
possible suspects to cause to difference between
simulators phase noise results i) noise transfer functions
ii) noise sources

Noise transfer functions seem to agree in both simulators.

Noise sources for devices do not seem to match.
I am still investigating though. I need to figure out
Spectre flicker and channel thermal noise settings; what
is default, and why etc. Once I get around Spectre
settings, then I will check longer length bias devices and sub-u channel length amplifer RF devices thermal noise difference in both simulators.

Title: Re: SpectreRF vs eldoRF phase noise simulation
Post by Paul on Jul 17^th, 2004, 1:26pm

Hi,

please keep the forum updated on your findings. I am also using both Spectre and Eldo and I am really interested in your comparison results.

Good luck

Paul